Re: Changing TERM TERM to HUP TERM

[Ole Tange]

On Mon, Mar 11, 2019 at 6:48 AM Gary Johnson <garyjohn@spocom.com> wrote:
>
> On 2019-03-10, Ole Tange wrote:
> > To stop GNU Parallel today you need to send TERM to make it stop
> > starting new jobs followed by another TERM to kill the running jobs.
:
> > What is your opinion?
>
> A change seems OK, but using HUP for this purpose doesn't seem like
> a good idea.  Its meaning is specified in the signal(7) man page and
> that doesn't seem to match your proposed usage.  How about using
> USR1 or USR2 instead?

USR1 and USR2 are already used:

LIST RUNNING JOBS
       If you want a list of the jobs currently running you can run:

         killall -USR1 parallel

       GNU parallel will then print the currently running jobs on
stderr (standard error).

           By sending GNU parallel SIGUSR2 you can toggle turning
on/off --progress on a running GNU parallel process.

If I look at the list of signals:

Signal     Value     Action   Comment
──────────────────────────────────────────────────────────────────────
SIGHUP        1       Term    Hangup detected on controlling terminal
                              or death of controlling process
SIGINT        2       Term    Interrupt from keyboard
SIGQUIT       3       Core    Quit from keyboard
SIGILL        4       Core    Illegal Instruction
SIGABRT       6       Core    Abort signal from abort(3)
SIGFPE        8       Core    Floating-point exception
SIGKILL       9       Term    Kill signal
SIGSEGV      11       Core    Invalid memory reference
SIGPIPE      13       Term    Broken pipe: write to pipe with no
                              readers; see pipe(7)
SIGALRM      14       Term    Timer signal from alarm(2)
SIGTERM      15       Term    Termination signal
SIGUSR1   30,10,16    Term    User-defined signal 1
SIGUSR2   31,12,17    Term    User-defined signal 2
SIGCHLD   20,17,18    Ign     Child stopped or terminated

SIGCONT   19,18,25    Cont    Continue if stopped
SIGSTOP   17,19,23    Stop    Stop process
SIGTSTP   18,20,24    Stop    Stop typed at terminal
SIGTTIN   21,21,26    Stop    Terminal input for background process
SIGTTOU   22,22,27    Stop    Terminal output for background process

The signals SIGKILL and SIGSTOP cannot be caught, blocked, or ignored.

Next the signals not in the POSIX.1-1990 standard but described in
SUSv2 and POSIX.1-2001.

Signal       Value     Action   Comment
────────────────────────────────────────────────────────────────────
SIGBUS      10,7,10     Core    Bus error (bad memory access)
SIGPOLL                 Term    Pollable event (Sys V).
                                Synonym for SIGIO
SIGPROF     27,27,29    Term    Profiling timer expired
SIGSYS      12,31,12    Core    Bad system call (SVr4);
                                see also seccomp(2)
SIGTRAP        5        Core    Trace/breakpoint trap
SIGURG      16,23,21    Ign     Urgent condition on socket (4.2BSD)
SIGVTALRM   26,26,28    Term    Virtual alarm clock (4.2BSD)
SIGXCPU     24,24,30    Core    CPU time limit exceeded (4.2BSD);
                                see setrlimit(2)
SIGXFSZ     25,25,31    Core    File size limit exceeded (4.2BSD);
                                see setrlimit(2)

I find it pretty clear that I cannot use a signal whose action is
'core': Those are not even close in meaning.

Neither are the ones with the action 'stop'. SIGURG is also not close
in meaning. SIGCHLD is used actively, so that rules that out.

So we are in reality down to the ones with action 'term'. And asking
GNU Parallel not to start more jobs is kind of asking it to terminate
soonish.

SIGALRM has a wrong ring to me.
SIGPIPE will no doubt be triggered by other means.
SIGINT is ctrl-C and should stay the same.
SIGKILL will kill GNU Parallel here and now - no cleanup. And it
cannot be caught.
SIGTERM will make GNU Parallel kill children and clean up.

That leaves SIGHUP.

I know SIGHUP is used in some daemons to say 're-read your config',
which is also not really the situation described in 'man 7 signal'.

I have now committed a version where SIGHUP is used. Can I ask you to
try to make it fail/do odd things that version 20190222 does not do?

If I do not hear any reports before Thursday, I will leave it in and
make it clear in next release that this is experimental (so that if we
discover bad behaviour, it will be rolled back).


/Ole